AI SEO Glossary

A bot that fetches pages to build an AI index or to answer a live query. Blocking one, on purpose or by accident, removes you from the answers it feeds.

Google's fully conversational search experience, where a generated answer, not a list of links, is the primary result. A distinct surface to be cited in.

Google's AI-written summary that can sit above the classic results for a query. It may answer the searcher outright, and being cited in it is a distinct outcome from ranking below it.

The practice of getting a brand found, understood, and cited inside AI-generated answers, not only ranked in a list of links. It shares SEO's technical base but is judged on citation rather than position.

How present, accurate, and well-cited a brand is across AI answers. The AI-era counterpart to search visibility, measured on whether assistants name and describe you correctly.

A single figure that rolls up how often and how favourably assistants surface a brand across a set of test questions. Useful as a trend line, not as an absolute.

Optimising so an answer engine names or cites you when it responds to a question directly. The near-synonym of GEO, framed around the answer rather than the generation.

Writing that leads with a direct, standalone answer before the context. It gives an assistant a passage it can quote without reconstructing your point from a whole article.

A reference to your page inside an AI answer, whether as a named source, a link, or a mention. Citation is the AI-era equivalent of the click you used to win.

Independent sources describing you the way you describe yourself. It is what moves a model from noting your claim to trusting it enough to repeat it.

The signals of credibility a source can carry. Assistants lean on the same signals when deciding whether a page is safe to cite on a given topic.

A distinct thing a model can identify and reason about, such as a company, person, or product. Being a clear, consistent entity is what lets an assistant place who you are.

The work of making your brand an unambiguous entity across your site, your structured data, and third-party sources, so a model can identify you with confidence.

How cleanly a self-contained answer can be lifted from your page into a response. Answer-first passages under clear headings are far easier to extract than buried prose.

Optimising to be one of the sources a generative system draws on when it writes an answer. The most common name for the discipline this handbook covers.

Tying a generated answer to retrieved sources rather than the model's memory. Grounded answers cite live pages, which is why retrievable, current content gets named.

A structured map of entities and the relationships between them. Consistent signals across the web feed it, and a clear place in it helps a model describe you accurately.

The kind of model behind ChatGPT, Gemini, and Claude. It predicts text, and when connected to live retrieval it can cite the web rather than rely only on training data.

The same work as AI SEO, named after the large language models doing the answering. A framing that stresses being retrievable and quotable by an LLM.

A proposed plain-text file at a site's root that summarises its key content for language models. Cheap to add, still early in adoption, and not a substitute for being crawlable.

The fixed list of buyer questions you run through assistants on a schedule to measure visibility. Keeping it stable is what turns one-off checks into a comparable trend.

When an assistant breaks one prompt into several sub-questions and answers the cluster. Covering the whole cluster, not just the head query, is what keeps you in the response.

The step where a system fetches candidate sources before writing an answer. If a crawler cannot reach and read your page, you are never in the retrieval set to begin with.

The pattern behind most cited answers: retrieve relevant sources, then generate a response grounded in them. It is why being retrievable matters as much as being good.

Structured data that labels what a page is about in a machine-readable form. It helps a model parse your content, though it does not by itself earn a citation.

Your slice of the citations across a set of buyer questions, measured against competitors. It reframes AI visibility from a yes-or-no into a competitive share.